A method and system for dispatching main network and distribution network in coordination considering fault risk in mountainous area

Abstract

The application discloses a kind of main distribution network collaborative scheduling method and system considering mountainous area fault risk. Meteorological, lightning activity, vegetation and line account and real-time operation data along the feeder in mountainous area are collected;The lightning stroke, mountain fire, tree barrier failure probability varying with meteorological and environmental quantity is established and calculated in section, and the probability model is verified and parameter updated with measured fault and meteorological data, to obtain the verified failure probability;Under the node voltage constraint, the available adjustment capacity of each adjustment unit is determined;According to the verified failure probability, the comprehensive risk index of each section is calculated and classified;In the upper limit of voltage constraint available capacity, with the redistribution weight determined by failure probability, the redistribution weight decreases with the increase of failure probability, the adjustment demand is weighted and distributed by risk weighted quadratic programming, to obtain the revised scheduling scheme.The external failure uncertainty of mountainous area is quantified explicitly, the probability model can be verified by measured data, so that the failure probability continuously drives the scheduling, while reducing the expected risk exposure, the standby margin of each adjustment unit is reserved, and the power supply reliability of long feeder in mountainous area is improved.

Description

Technical Field

[0001] This invention relates to the field of power dispatching technology, and in particular to a method and system for coordinated dispatching of main and distribution networks that takes into account the risk of faults in mountainous areas. Background Technology

[0002] Long feeder lines in mountainous areas, exemplified by 10kV main-distribution interconnection feeder lines traversing forest regions, are characterized by their length, dense vegetation, and complex terrain. External environmental factors significantly increase the probability of faults compared to urban power grids, primarily including lightning strike tripping, wildfire flashovers, and tree-related faults caused by tree growth. All three types of faults are strongly driven by weather and seasonal factors and are unevenly distributed along different sections of the feeder. The long feeder lines at the end, traversing dense forests and areas prone to lightning strikes, show significantly higher fault exposure compared to the open sections on the outgoing side. Sectional switches or automatic reclosers are typically installed between feeder sections. After a fault is isolated, only the load in that section is lost; downstream sections are restored to power via the main-distribution interconnection switch or the distributed power source at the end.

[0003] Existing coordinated dispatching of main and distribution networks mostly adopts deterministic models, or only considers internal uncertainties such as the fluctuation of distributed renewable energy output, without considering the aforementioned external uncertainties such as faults. Even if some methods conduct risk assessments, they only treat risk as a switch to trigger dispatching. The result of the dispatching correction, i.e., how much regulation each unit undertakes, is not affected by the fault probability value. High-risk fault sections and low-risk fault sections are treated equally. Once a fault occurs in a high-risk section, the load loss in that section is greater.

[0004] In addition, voltage fluctuations are significant when power is injected at the end of long feeders in mountainous areas, and the actual available regulation capacity of the distribution network is often lower than the sum of the rated capacities of each unit. Furthermore, the long feeder section at the end is often both a high-risk fault section and a voltage bottleneck section. Summarizing and evenly distributing the load according to the rated capacity will overestimate the capacity and put the load on the weak section.

[0005] Therefore, a method and system for coordinated scheduling of main and distribution networks that takes into account the risk of faults in mountainous areas is needed. Summary of the Invention

[0006] To address the problems existing in the prior art, this invention provides a method and system for coordinated scheduling of main and distribution networks that considers the risk of faults in mountainous areas. The specific technical solution is as follows: A method for coordinated scheduling of main and distribution networks considering fault risks in mountainous areas includes: S1. Collect meteorological data, lightning activity data, vegetation and line ledger data, and real-time operation data of the power distribution network along the mountain feeder line; S2. Based on the collected data, establish and calculate the probability of external environment-induced failures for each section of the feeder, wherein the external environment-induced failures include at least one of lightning strike failure probability, wildfire failure, and tree obstacle failure. S3. The reliability of the fault probability model is verified by using measured fault statistics and meteorological data. When the verification deviation exceeds the set threshold, the corresponding model parameters are updated to obtain the verified fault probability. S4. Under the constraint of node voltage, determine the available regulation capacity of each regulation unit in the distribution network, and obtain the regulation gap based on the main grid dispatch instructions and the sum of each available regulation capacity; S5. Based on the verified fault probability, calculate the comprehensive risk index of each section of the feeder, summarize the calculations to obtain the comprehensive risk index of the feeder, and classify the feeder index. S6. If the regulation gap is equal to 0, then within the upper limit of the available regulation capacity of the voltage constraints of each regulation unit, the regulation demand is weighted and allocated according to the redistribution weight determined by the fault probability after verification, so as to obtain the corrected scheduling scheme. If the adjustment gap is greater than 0, the adjustment amount of each adjustment unit is set to its own upper limit of available adjustment capacity to form a scheduling scheme.

[0007] Preferably, step S2 includes: The probability of lightning strikes increases with the increase of ground flash density and the number of lightning strikes on the line in the section; the probability of wildfires increases with the increase of the fire risk index calculated from air temperature, relative humidity, wind speed and drought factor; the probability of tree obstruction decreases with the increase of the clearance margin between the conductor and the vegetation corridor.

[0008] Preferably, step S3 includes: The relative deviation between the predicted failure frequency and the measured failure frequency of the failure probability model is calculated, and the Brier score between the predicted probability and the actual occurrence of the event is calculated. When the relative deviation and the Brier score do not exceed their respective set thresholds, the current model is adopted and marked as highly reliable; otherwise, the model parameters are updated recursively.

[0009] Preferably, the model parameters include the tripping conversion coefficient of the lightning strike model, the regression coefficient of the wildfire model, and the vegetation growth rate and feature margin of the tree barrier model, and the verified fault probability is recalculated with the updated parameters.

[0010] Preferably, step S4 includes: Linearize the power flow equations at the current operating point to obtain the voltage-power sensitivity of each unit's injected power to the critical node voltage; The maximum injectable power of the unit under voltage constraints is obtained by dividing the difference between the node voltage limit and the current voltage of the critical node by the voltage-power sensitivity. The smaller value between the maximum injectable power and the rated adjustable capability of the unit is taken as the available adjustable capability of the unit; The difference is obtained by subtracting the sum of the available adjustment capabilities of each unit from the main network scheduling command. When the difference is greater than 0, the difference is the adjustment gap; when the difference is less than or equal to 0, the adjustment gap is 0.

[0011] Preferably, step S5 includes: Calculate the comprehensive risk score for each section; The feeder comprehensive risk index is calculated by summing the product of the comprehensive risk score of each section and the load power lost after the section is isolated by the sectionalizing switch on each section; and the feeder comprehensive risk index is divided into three levels: high, medium and low according to a set threshold.

[0012] Preferably, the set threshold is determined based on the quantile of the historical power loss distribution of the feeder, or based on the power supply reliability target and the acceptable expected power loss level.

[0013] Preferably, in step S6, when the adjustment gap is equal to 0, the following steps are included: The objective is to minimize the sum of the squares of the adjustment amounts of each adjustment unit after verification by the fault probability of its respective section, and the constraint is that the sum of the adjustment amounts of each unit is equal to the main grid dispatching instruction, and the adjustment amounts of each unit do not exceed the upper limit of the available adjustment capacity of the voltage constraint of that unit. When all constraints are not saturated, the solution is that each unit shares the adjustment demand proportionally according to the redistribution weight. The redistribution weight is the reciprocal of the sum of the fault probability of the section where the unit is located after verification and a regularization parameter. When the allocation of a certain unit reaches the upper limit of its voltage constraint available adjustment capability, the unit is fixed at the upper limit and deducted from the adjustment demand to be allocated. Then the remaining units are solved again, and the cycle is repeated until no unit exceeds the limit, and the corrected scheduling scheme is obtained.

[0014] Preferably, step S6 further includes: The feeder comprehensive risk index, the available regulation capacity of each regulation unit, and the regulation gap are reported to the main grid. Based on this, the main grid regenerates the next cycle dispatch instruction and issues it to the distribution network, forming an iterative collaboration between the main grid and the distribution network.

[0015] A main and distribution network coordinated scheduling method that considers the fault risk in mountainous areas, applied to the above-mentioned method, includes a data acquisition module, a fault probability modeling module, a meteorological verification and parameter update module, a power flow and regulation capacity assessment module, a comprehensive risk assessment module, and a deterministic scheduling correction and decision support module. The data acquisition module is used to collect meteorological data, lightning activity data, vegetation and line ledger data, and real-time operation data of the power distribution network. The failure probability modeling module is used to calculate the failure probability of each partition segment; The meteorological verification and parameter update module is used to calculate the relative deviation between the predicted fault frequency and the measured fault frequency or to calculate the Brier score between the predicted probability and the actual occurrence of the event, and compare it with the respective set thresholds to determine whether to perform an update verification. The power flow and regulation capacity assessment module is used to calculate the available regulation capacity of each regulation unit in the distribution network to obtain the regulation gap; The comprehensive risk assessment module is used to calculate the risk index and classification based on external environment data; The deterministic scheduling correction and decision support module is used to output the corrected scheduling scheme and risk warning.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: In this invention, the uncertainty of external faults in mountainous areas can be explicitly quantified, closed-loop verification can be performed using measured data, and the fault probability value can be continuously driven and allocated within the voltage feasible domain after verification, thereby reducing the exposure of expected risks while maintaining the reserve margin of each unit. Attached Figure Description

[0017] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0018] Figure 1 This is a flowchart of the method of the present invention.

[0019] Figure 2 This is a schematic diagram of the three types of fault scenarios and probability models for long feeder lines in mountainous areas according to the present invention.

[0020] Figure 3 This is a closed-loop diagram of meteorological verification and parameter update for this invention.

[0021] Figure 4 This is a topology diagram of a segmented long feeder line in the Zhongshan area, as described in this invention.

[0022] Figure 5 System module structure diagram in this invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0025] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0026] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0027] Example 1 Please see Figure 1 This embodiment provides a method for coordinated scheduling of main and distribution networks considering the risk of faults in mountainous areas, specifically including the following steps: S1. Collect meteorological data, lightning activity data, vegetation and line ledger data, and real-time operation data of the power distribution network along the mountain feeder line; S2. Based on the collected data, establish and calculate the probability of external environment-induced failures for each section of the feeder, wherein the external environment-induced failures include at least one of lightning strike failure probability, wildfire failure, and tree obstacle failure. Among them, the models for the probability of failure caused by the external environment are all monotonic mapping models that map meteorological and environmental quantities to the interval between 0 and 1, as detailed below: (1) The probability of lightning strike failure increases with the increase of ground flash density and the number of lightning strikes on the line in the section. The specific calculation is as follows: Regarding the probability of lightning strike failure, the number of lightning strikes on the line in the section is: Lightning trip rate: Evaluation period The probability of an internal lightning strike failure is: In the formula For ground flash density, The average height of the conductor above the ground. The horizontal spacing of the conductors. For segment length, This is the trip conversion factor. The assessment period is as follows. Because the amplitude of lightning current in the vast majority of cases exceeds the critical current... ,in This is 50% of the flashover voltage of the insulator. Since the conductor's impedance is constant, the probability of a strike-and-flashover is approximately 1. Take the initial value determined by the power frequency follow current arc-building rate and update it by calibration in step S3.

[0028] It should also be noted that the tripping conversion coefficient... The arc-building rate characterizes the proportion of power frequency follow current that establishes a stable arc after a lightning flashover and progresses from flashover to sustained tripping. Its initial value is determined by the power frequency follow current arc-building rate of the line. The arc-building rate increases with the power frequency voltage gradient of the insulator string and can be determined empirically. Give, This represents the power frequency recovery voltage gradient along the arc path. For 10kV and 35kV overhead distribution lines, this initial value varies depending on the neutral grounding method: in systems where the neutral point is ungrounded or grounded via an arc suppression coil, single-phase ground faults have small power frequency follow currents and are mostly self-extinguishing. Initial values ​​are taken as 0.10 to 0.30; in systems where the neutral point is grounded through a small resistor, the fault current is larger and the arc-building rate is higher. The initial value is set to 0.30 to 0.60. The above initial value is only used as the starting point of the iteration, and its accuracy does not affect the convergence of this method. It is updated to a stable value by step S3 using the measured tripping statistics calibration.

[0029] (2) The probability of wildfire failure increases with the increase of the fire risk index calculated from air temperature, relative humidity, wind speed and drought factor. The specific calculation is as follows: Regarding the probability of wildfire failure, the fire risk index is calculated based on meteorological measurements: The probability of wildfire failure is: In the formula For temperature, Relative humidity, For wind speed, As a drought factor, For the density of vegetation in the corridor, , , Let be the regression coefficients fitted from historical wildfire power outage records, and take . , To ensure monotony.

[0030] It should be noted that the regression coefficients... , , The data was determined by fitting the target area's historical wildfire trip records using maximum likelihood logistic regression. The samples consisted of binary labels indicating whether wildfire trips occurred during various fire weather events in the area, along with their corresponding fire risk indices and corridor vegetation densities. The goodness of fit was assessed using the McFadden pseudo-fit. Or AIC evaluation, and requirements , The estimated value is significantly positive, causing the probability of wildfire failure to increase monotonically with the fire risk index and vegetation density. The regression coefficient varies with the vegetation and fire risk climate characteristics of the region. When this method is used in different climate zones, it needs to be refitted with historical samples of that region. To ensure the robustness of the coefficient estimation, the refitted samples should cover multiple fire seasons and have at least several dozen positive samples of wildfire tripping.

[0031] (3) The probability of tree obstruction failure decreases as the clearance margin between the conductor and the vegetation corridor increases. The specific calculation is as follows: Regarding the probability of tree obstruction failure, the real-time clearance margin is: The sag increment is: conductor temperature From the heat balance equation: Solve for the probability of tree obstacle failure: In the formula This is the initial headroom. For vegetation growth rate, The time since the last cleaning, Sag-temperature coefficient For reference temperature, For characteristic margin, For conductor load current, The AC resistance of the wire. , , The heat absorption from solar radiation, the heat dissipation from convection, and the heat dissipation from radiation are respectively defined, and the following conventions are made. hour .

[0032] S3. The reliability of the fault probability model is verified using measured fault statistics and meteorological data. When the verification deviation exceeds a set threshold, the corresponding model parameters are updated to obtain the verified fault probability; including: The relative deviation between the predicted fault frequency and the measured fault frequency of the fault probability model is calculated, as well as the Brier score between the predicted probability and the actual occurrence of the event. When both the relative deviation and the Brier score do not exceed their respective set thresholds, the current model is adopted and marked as highly reliable; otherwise, the model parameters are recursively updated. The model parameters include the tripping conversion coefficient of the lightning strike model, the regression coefficient of the wildfire model, and the vegetation growth rate and feature margin of the tree barrier model. The validated fault probability is recalculated using the updated parameters.

[0033] Specifically: Calculate and predict the frequency of failures Compared with the measured fault frequency Relative deviation: Calculate the Brier score for probability classes. In the formula For the first Next prediction probability. For the first This event occurred as an indicator value. The number of samples used for verification. When and If none of the parameters exceed their respective thresholds, the model is adopted and marked as highly reliable; otherwise, the model parameters are updated recursively.

[0034] For example, the tripping conversion factor in the lightning strike model is calculated as follows: renew, The step size coefficients are those taking values ​​between 0 and 1, representing the regression coefficients of the wildfire model. Tree barrier model parameters , Each is updated separately using least-squares recursion based on its measured samples. The relative deviation threshold... Brier score threshold This reflects the tolerance for the reliability of the failure probability model, and can be either a fixed threshold or adaptively adjusted according to the scenario. A threshold that is too strict will frequently trigger false parameter updates due to normal statistical fluctuations in the measured frequency, causing parameter oscillations; a threshold that is too wide will render the verification ineffective, and the failure model will not be corrected. When using a fixed threshold, Take a value between 0.15 and 0.25. The threshold value ranges from 0.05 to 0.15, corresponding to the width of the confidence interval for the measured frequency estimate at a given significance level. When using an adaptive threshold, the threshold is adjusted based on the following factors: the higher the feeder voltage level and the stricter the power supply reliability requirements, the smaller the threshold should be to trigger updates earlier; the smaller the measured sample size and the greater the statistical fluctuations in the frequency estimate, the wider the threshold should be. In form, and For the baseline threshold and the baseline sample size, The threshold is set to the number of measured samples to correspond to the statistical significance level of the measured frequency estimate. The higher the frequency of lightning activity or wildfires in a region and the more sufficient the available samples, the tighter the threshold can be to more sensitively correct the failure model.

[0035] S4. Under node voltage constraints, determine the available regulation capacity of each regulating unit in the distribution network, and obtain the regulation gap based on the main grid dispatch instructions and the sum of the available regulation capacities, including: Linearizing the power flow equations at the current operating point yields the voltage-power sensitivity of each unit's injected power to the critical node voltage; specifically, regarding the adjustable capability of voltage constraints, the unit... Voltage-power sensitivity: The maximum injectable power of the cell under voltage constraints is obtained by dividing the difference between the node voltage upper limit and the current voltage of the critical node by the voltage-power sensitivity; specifically, the maximum injectable power under voltage constraints is obtained by linearizing the power flow equations at the operating point: The smaller value between the maximum injectable power and the rated adjustable capability of the unit is taken as the available adjustable capability of the unit; specifically: the available adjustable capability is taken as... The regulation gap is obtained by subtracting the sum of the available regulation capabilities of each unit from the main network scheduling instructions. The regulation gap is... In the formula This is the upper limit of the node voltage. The current voltage of the critical node. The rated adjustment capacity of the unit, Main network scheduling instructions, For unit The injection power.

[0036] S5. Based on the verified fault probability, calculate the comprehensive risk index for each section of the feeder, summarize the calculated comprehensive risk index of the feeder, and classify the feeder index; including: Calculate the comprehensive risk score for each section; specifically: In the formula Traverse the segments isolated by the segment switches. For section No. Class failure probability, This represents the load power lost in this section after isolation.

[0037] The feeder comprehensive risk index is calculated by summing the product of the comprehensive risk score of each section and the load power lost after isolation by the sectionalizing switch in each section. Specifically: The feeder comprehensive risk index is divided into three levels: high, medium, and low, according to a set threshold. The set threshold is determined based on the quantile of the historical power loss distribution of the feeder or based on the power supply reliability target and the acceptable expected power loss level.

[0038] S6. If the regulation gap is equal to 0, then within the upper limit of the available regulation capacity of the voltage constraints of each regulation unit, the regulation demand is weighted and allocated according to the redistribution weight determined by the fault probability after verification, so as to obtain the corrected scheduling scheme. If the adjustment gap is greater than 0, the adjustment amount of each adjustment unit is set to its respective upper limit of available adjustment capacity to form a scheduling scheme. When the adjustment gap is equal to 0, it includes: The objective is to minimize the sum of the squares of the adjustment amounts of each adjustment unit after verification by the fault probability of its respective section, and the constraint is that the sum of the adjustment amounts of each unit is equal to the main grid dispatching instruction, and the adjustment amounts of each unit do not exceed the upper limit of the available adjustment capacity of the voltage constraint of that unit. When all constraints are not saturated, the solution is that each unit shares the adjustment demand proportionally according to the redistribution weight. The redistribution weight is the reciprocal of the sum of the fault probability of the section where the unit is located after verification and a regularization parameter. When the allocation of a unit reaches the upper limit of its voltage constraint-available regulation capacity, the unit is fixed at that upper limit and deducted from the regulation demand to be allocated. Then, the remaining units are solved again, and the process is repeated until no unit exceeds the limit, resulting in a revised scheduling scheme. The specific calculation method is as follows: In the risk-weighted quadratic programming context, the weighted allocation is the solution to the following quadratic programming problem: Its constraints are and When all constraints are unsaturated, the analytical solution is: The redistribution weights are: In the formula For unit The fault probability of the section after verification in step S3. To prevent excessive concentration of allocation in low-risk units, the regularization parameter is taken to be on the same order of magnitude as the minimum segment failure probability; when the allocation of a unit reaches its upper limit of available adjustment capacity, it is fixed at the upper limit and then... After deducting from the middle, the remaining units are solved again using the above formula, which only includes the sum of their weights, and the cycle continues until no unit exceeds the limit.

[0039] The regularization parameter Minimum segment failure probability Same order of magnitude, take proportionality coefficient Take 0.5 to 2. The effect on the allocation result is monotonic: The larger the value, the more equal the redistribution weights become, the closer the allocation is to an equal distribution based on available adjustment capacity, and the lower the risk differentiation. The smaller the value, the greater the weight of low-risk units, the more they are centrally allocated, and the easier it is for their reserve margin to be exhausted. In engineering terms... It can be determined by working backward from the acceptable lower bound of the reserve margin: given the minimum reserve margin required to retain a non-voltage bottleneck unit, substitute it into the unsaturated analytical solution. By ensuring that the allocation of this unit does not exceed the difference between its available adjustment capacity and this margin, the solution can be obtained. The lower bound can be determined by cross-validation, which can be used to select the optimal balance between expected risk exposure and reserve margin in historical scenarios. .

[0040] The following examples illustrate several scenarios to further illustrate the invention: Scenario 1: High lightning and fire risk in summer for long feeder lines in mountainous areas, with voltage bottleneck at the end. A complete demonstration of steps S1 to S6 in a typical scenario. The scenario and object are: a 10kV long feeder line in a mountainous area divided into three sections. The near-end section A is located on the outgoing side with an open corridor; the middle section B partially passes through forest and includes a long, heavily loaded branch line, with weak electrical connections and a small voltage margin; the end section C is located at the end of the long line, passing through dense forest, in a lightning-prone area, and is the voltage bottleneck of the long line. Sectional switches isolate each section. The loads of each section are 1.2MW, 1.0MW, and 0.8MW, respectively, totaling 3.0MW. Three regulating units U1, U2, and U3 are connected to sections A, B, and C, respectively, with rated capacities of 0.7MW, 0.4MW, and 0.6MW. The node voltage upper limit is... pu, mainnet command MW.

[0041] Step S1: Collect meteorological and lightning activity data, vegetation and line records, and real-time operation data of the distribution network along the feeder line as input for subsequent steps.

[0042] Step S2: Calculate the probability of three types of faults by segment, referring to... Figure 2 Taking the final segment C as an example: Figure 2 shows the meteorological and line input quantities required for the on-site scenarios of faults caused by three types of external environmental factors, namely lightning strikes, wildfires, and tree obstructions, in the three sections of the long feeder line in the mountainous area: the open section A at the near end, the middle section B, and the forested section C at the end.

[0043] Regarding the probability of lightning strike failure, the ground flash density in section C... Times / (km²·year), average height of conductor above ground m, horizontal spacing m, segment length km. The number of lightning strikes is calculated according to the IEEE 1410 lightning strike width formula, which is well known in the field: Once per year.

[0044] Pick kV, Ω, critical lightning current kA, the median lightning current is about 31kA, much greater than The probability of a flashover upon impact is approximately 1. An initial value for the arc-forming rate of a 10kV line is taken. This initial value is updated by S3 calibration: Once per year.

[0045] Pick Days, or 30 / 365 years: The relative risk index for lightning strike failures is approximately 4.07%.

[0046] Regarding the probability of wildfire failures, the weather conditions during the fire season in section C are: ℃ %, km / h , The fire risk index is calculated according to McArthur Mark 5 formula, which is well known in the field: The area is classified as high fire risk. The probability of wildfire failure is calculated using logistic regression, taking... , , The linear term = −5.0 + 0.25 × 12.1 + 1.0 × 0.7 = −1.275: This means the relative risk index for wildfire failure conditions is approximately 21.8%.

[0047] Regarding the probability of tree obstacle failure, the parameter for section C is: m、 m / year Year. Under heavy load current. A. At an ambient temperature of 35℃, thermal equilibrium is achieved according to IEEE 738. Determine the temperature of the conductor ℃, take ℃ m / ℃, then m; Real-time clearance margin m; take m, then The relative risk index for tree obstacle failure is approximately 18.9%.

[0048] Regarding the comprehensive risk score for a section, section C is determined by a combination of three types of fault risk indices. Sections A and B, calculated using the same model under their respective relatively mild weather and vegetation conditions, yielded the following results: , .Should The comprehensive risk score used for classification and weighting is not the actual concurrent probability of the three types of failures.

[0049] The above-mentioned segmented fault probability calculation is a specific implementation of step S2. The probabilities of the three types of faults, namely lightning strike, wildfire, and tree obstruction, are all calculated in segments according to the monotonic model that maps meteorological and environmental quantities to the interval of 0 to 1.

[0050] Step S3, meteorological verification and parameter update. Referring to Figure 3, taking the lightning strike model for section C as an example, the prediction frequency... Frequency per year, actual measured conversion frequency Once per year.

[0051] Figure 3 The meteorological verification and parameter update closed loop in step S3 is shown. The relative deviation or Brier score is obtained by comparing the predicted fault frequency with the measured fault frequency. If the deviation exceeds the threshold, the corresponding model parameters are updated recursively. After the update, the prediction is re-established, forming a closed loop of verification, update and re-prediction.

[0052] deviation Trigger update; press Update, retrieve ,have to After the update The verification passed. The verification changes the downstream output, updating the lightning strike probability. Section C Comprehensive Risk Score That is, it decreased from 0.392 to 0.387, and this change was adjusted by weights. Input the quadratic programming from step S6 and change the adjustment amount of U3. Wildfire and tree barrier model parameters. , , Each is updated separately using least-squares recursion based on its measured samples, and expressed as Brier scores. Verification, its threshold .

[0053] The above verification and update uses the relative deviation between the predicted fault frequency and the measured fault frequency or the probability class Brier score as the verification quantity. When the deviation exceeds the threshold, the corresponding model parameters are recursively updated according to the step size coefficient.

[0054] Step S4, available adjustable capability under voltage constraint, refer to Figure 4 The upper limit of the voltage of each unit is taken as pu, current critical node voltage The voltage pu is caused by the reverse feed from the distributed power source.

[0055] Figure 4 shows the segmented topology of the example feeder. From the outgoing side, the segments are A (low-risk section near the end), B (middle section), and C (high-risk section at the end with a voltage bottleneck). The three segments are connected to regulating units U1, U2, and U3, respectively. Adjacent segments are isolated by sectionalizing switches. The voltage critical node is marked at the end. The relative positions and electrical connections of each segment, regulating unit, sectionalizing switch, and voltage critical node are shown.

[0056] U1 is close to section A, with a wide corridor, short electrical distance, and sufficient voltage margin. Much greater than the rated value, MW.

[0057] U2 is connected to section B in the middle section, which has weak electrical connections and high voltage-power sensitivity. pu / MW, then MW, The rated capacity is 0.4MW, but due to voltage exceeding the limit, the actual capacity is only 0.25MW.

[0058] U3 is connected to the terminal section C, which is a bottleneck for long-line voltage. pu / MW, then MW, MW.

[0059] Always available MW, MW is sufficient and there is no gap. The key to scheduling is to allocate 1.0MW among the three units.

[0060] The specific implementation of the above-mentioned available regulation capacity calculation step S4 involves linearizing the power flow equation at the current operating point to obtain the voltage-power sensitivity of each unit to the critical node, back-calculating the maximum injectable power constrained by voltage based on the upper voltage limit, and taking the smaller value between it and the rated regulation capacity.

[0061] Step S5: Divide the overall risk index by segment. Calculate the overall risk score for each segment. , , The load loss in each section after isolation , , MW, the comprehensive risk index is taken as the expected power loss: MW.

[0062] Because each section is isolated, the fault only affects the load of that section, and the load is not counted repeatedly for each section. The total load is 3.0MW, physically self-consistent. The grading thresholds are determined based on historical load shedding quantiles; in this example, high risk is greater than 0.5MW, medium risk is 0.2 to 0.5MW, and low risk is less than 0.2MW. MW was deemed high-risk, triggering a risk-weighted redistribution.

[0063] The above-mentioned comprehensive risk index calculation is the specific implementation of step S5. The comprehensive risk score of a section is 1 minus the product of the probabilities that all types of faults in that section will not occur. The comprehensive risk index of the feeder is the sum of the products of the risk scores of each section and the load lost after isolation. Step S6, risk-weighted quadratic programming redistribution, see Figures 1 and 4. Available. , , MW, probability of failure after verification , , Establish risk-weighted secondary programming. , constraint is , ,Pick First, find analytical solutions based on the unsaturated conditions at each upper limit. ,have to , , , , tentative solution , , MW; among which If MW exceeds its voltage constraint upper limit of 0.25MW, i.e., it exceeds the limit, triggering an active-set re-resolution. U2 is fixed at its voltage upper limit of 0.25MW, and the remaining elements are... Resolve the remaining 0.75MW to get MW MW, final allocation , , MW, summed to 1.000MW; U2 is exactly at its voltage constraint upper limit, and U1 and U3 retain reserve margins of 0.088MW and 0.262MW respectively.

[0064] The above weighted allocation is the specific implementation of step S6. Solve the quadratic programming of the sum of squares of the adjustment amount of each unit weighted by the failure probability after verification, and fix the units that have reached the upper limit of available adjustment capacity in the form of an effective set and then solve it again; report the feeder comprehensive risk index, the available adjustment capacity of each unit and the adjustment gap to the main network to form the main-feeder iterative collaboration.

[0065] In terms of technical effectiveness, the expected risk exposure is based on the full feeder diameter. By measurement, the present invention obtains The risk-blind allocation, based on the proportion of available capacity, yields approximately 0.170MW, while the greedy allocation, sequentially filling in the risk-blind areas, yields approximately 0.092MW. Compared to the risk-blind proportional allocation, this method reduces the expected risk exposure of the entire feeder from 0.170MW to 0.122MW, a reduction of approximately 28.0%. Compared to the greedy allocation, this method results in a slightly higher exposure of approximately 0.030MW. However, the greedy allocation fills the lowest-risk unit U1 to zero reserve margin, while this method retains reserve margins for non-voltage bottleneck units U1 and U3, with failure probability values ​​continuously entering the allocation. This method achieves a compromise between reducing expected exposure and retaining reserve margins within the voltage feasible region, a compromise that neither the greedy nor proportional allocation can simultaneously achieve.

[0066] Regarding the coordination between primary and secondary suppliers, a comprehensive risk index will be used. Each unit's available capacity and adjustment gaps are reported to the main network. Based on this, the main network generates the next cycle's instructions and issues them to the distribution network, forming iterative coordination. The continuous impact of verification on the output can be further explained in step S3. When the value is updated from 0.392 to 0.387, U2 remains fixed at 0.25MW, and the solution is obtained again. MW MW indicates that the measured verification amplitude continuously changes the scheduling output through weighting, rather than being triggered by a switch. Therefore, the fault probability mechanism, voltage constraints, and scheduling output are strongly coupled and inseparable.

[0067] Scenario 2: Open corridor, fire-prone middle section, and no voltage bottleneck. Another application scenario of this invention illustrates that, in the absence of voltage constraint binding, the allocation degenerates into a purely risk-weighted analytical solution, adapting to different risk space orders. The scenario involves a relatively open mountain feeder line divided into three sections: D, E, and F. Section E traverses a dense thicket of grass and shrubs, posing the highest fire risk. Three regulating units, V1, V2, and V3, are connected to the three sections respectively, with rated capacities of 0.6MW, 0.5MW, and 0.5MW. The loads of each section are 1.0MW, 0.9MW, and 0.8MW, respectively. The main grid command... MW.

[0068] Step S2: Using the same probability model for lightning strikes, wildfires, and tree obstructions as in Example 1, the comprehensive risk score for the three sections is calculated under the meteorological and line parameters of this scenario. , , The highest risk is in the middle section E, and its risk spatial order is different from that in Example 1.

[0069] Step S4: Because the corridor is spacious and the voltage margin of all three units is sufficient, the maximum injectable power of each unit under voltage constraints is greater than its rated power. Therefore, the available adjustable capacity is equal to the rated capacity, i.e. , , MW, total available 1.6MW, greater than the command, no gap.

[0070] Step S5, the load loss of each segment after isolation , , MW, Comprehensive Risk Index MW is classified as medium risk according to the same grading threshold.

[0071] Step S6, take Weight , , , Since none of the available capabilities are bound, the analytical solution is the final allocation. , , MW, summed to 1.000MW, no active-set re-solution is needed, the middle section E with the highest fire risk only gets 0.153MW because of the smallest weight.

[0072] In terms of technical effectiveness, this invention aims to expose risks. MW; the risk blind is allocated according to the proportion of available capacity as 0.375 / 0.313 / 0.313MW, with an expected exposure of approximately 0.148MW. This method reduces the expected risk exposure from 0.148MW to 0.111MW, a reduction of approximately 24.6%, indicating that in scenarios with different risk space orders and no voltage binding, this method still continuously drives the allocation based on the failure probability value, thereby reducing the expected risk exposure.

[0073] Scenario 3: Extreme weather and insufficient available capacity create a regulatory gap. This scenario illustrates another application case, demonstrating how, when the available regulation capacity of the distribution network is insufficient to meet the main grid's commands, the present invention calculates the regulation gap in step S4 and reports it to the main grid in step S6, thus forming a main-distribution coordination mechanism. The scenario is an extreme day with strong convection and high fire risk, resulting in severe voltage degradation at the end of the distribution network and high load demand. The feeder is divided into three sections: G, H, and J, with three regulation units rated at 0.6MW, 0.4MW, and 0.5MW respectively. The main grid commands... MW.

[0074] Step S2: Calculate the comprehensive risk score using the same model. , , Terminal J has the highest risk.

[0075] In step S4, due to severe deterioration of the terminal voltage, the maximum injectable power of the three cells under voltage constraints is significantly reduced, and the available adjustable power is... , MW, always available MW, less than the main network command of 1.3MW, adjustment gap. MW.

[0076] Steps S5 and S6: Load loss after isolation in each segment , , MW, Comprehensive Risk Index MW is classified as high-risk. This is due to the equality constraints of quadratic programming within the available domain. This is not feasible with a total available capacity of 0.95MW. The distribution network should respond to the maximum capacity available in each unit, i.e. , , The system calculates the MW (MW) and reports the 0.35MW adjustment gap, along with the comprehensive risk index and the available capacity of each unit, to the main grid. Based on this, the main grid regenerates the next cycle's dispatch instructions, prioritizing the use of adjacent feeders or adjusting the main grid power flow to cover the shortfall at the high-risk end J, and issues the instructions to the distribution network, forming an iterative coordination between the main and distribution networks. This scenario demonstrates that even with limited capacity, the present invention can still quantify the gap and drive main-distribution coordination and risk warning with risk information.

[0077] Example 2 Please see Figure 5 This embodiment discloses a main and distribution network coordinated dispatching system considering fault risks in mountainous areas, including a data acquisition module, a fault probability modeling module, a meteorological verification and parameter update module, a power flow and regulation capacity assessment module, a comprehensive risk assessment module, and a deterministic dispatch correction and decision support module. The data acquisition module collects meteorological data, lightning activity data, vegetation and line ledger data, and real-time operation data of the distribution network. The fault probability modeling module calculates the fault probability of each section. The meteorological verification and parameter update module calculates the relative deviation between predicted fault frequencies and measured fault frequencies, or calculates the Brier score between predicted probabilities and actual event occurrences, and compares it with their respective set thresholds to determine whether to perform an update verification. The power flow and regulation capacity assessment module calculates the available regulation capacity of each regulation unit in the distribution network to obtain the regulation gap. The comprehensive risk assessment module calculates the risk index and classification based on external environmental data. The deterministic dispatch correction and decision support module outputs the corrected dispatching scheme and risk warnings.

[0078] Specifically, the data acquisition module provides input data to the meteorological verification and parameter update module, jointly supporting the development of fault probability modeling. The probability modeling results support the calculation of the power flow and regulation capacity assessment module, calculating the probability of external environmental faults in each segment. The calculation results of the meteorological verification and parameter update module and the power flow and regulation capacity assessment module support the analysis and processing of the comprehensive risk assessment module, calculating and classifying the feeder comprehensive risk index. The deterministic dispatch correction and decision support module comprehensively utilizes the available regulation capacity and regulation gap data of each unit under deterministic voltage constraints from the power flow and regulation capacity assessment module, as well as the R=∑P·C and classification results from the comprehensive risk assessment module. Using a redistribution weight that decreases with increasing fault probability, it weights and allocates regulation demand within the upper limit of available regulation capacity under voltage constraints, obtaining a corrected dispatch scheme and outputting risk warnings. Through this system, the coordinated dispatch capability of the main and distribution networks of the prefecture-level power grid is improved, enabling fault probability to continuously drive dispatch, reducing expected risk exposure while maintaining reserve margins for each regulation unit, and improving the power supply reliability of long feeders in mountainous areas.

[0079] Those skilled in the art will recognize that the units of the various examples described in connection with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of the invention.

[0080] In the embodiments provided by the present invention, it should be understood that the division of units is only a logical functional division. In actual implementation, there may be other division methods, such as multiple units can be combined into one unit, one unit can be split into multiple units, or some features can be ignored.

[0081] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0082] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0083] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A method for coordinated scheduling of main and distribution networks considering fault risks in mountainous areas, characterized in that, include: S1. Collect meteorological data, lightning activity data, vegetation and line ledger data, and real-time operation data of the power distribution network along the mountain feeder line; S2. Based on the collected data, establish and calculate the probability of external environment-induced failures for each section of the feeder, wherein the external environment-induced failures include at least one of lightning strike failure probability, wildfire failure, and tree obstacle failure. S3. The reliability of the fault probability model is verified by using measured fault statistics and meteorological data. When the verification deviation exceeds the set threshold, the corresponding model parameters are updated to obtain the verified fault probability. S4. Under the constraint of node voltage, determine the available regulation capacity of each regulation unit in the distribution network, and obtain the regulation gap based on the main grid dispatch instructions and the sum of each available regulation capacity; S5. Based on the verified fault probability, calculate the comprehensive risk index of each section of the feeder, summarize the calculations to obtain the comprehensive risk index of the feeder, and classify the feeder index. S6. If the regulation gap is equal to 0, then within the upper limit of the available regulation capacity of the voltage constraints of each regulation unit, the regulation demand is weighted and allocated according to the redistribution weight determined by the fault probability after verification, so as to obtain the corrected scheduling scheme. If the adjustment gap is greater than 0, the adjustment amount of each adjustment unit is set to its own upper limit of available adjustment capacity to form a scheduling scheme.

2. The main and distribution network coordinated scheduling method considering fault risk in mountainous areas according to claim 1, characterized in that, in, Step S2 above also includes: The probability of lightning strikes increases with the increase of ground flash density and the number of lightning strikes on the line in the section; the probability of wildfires increases with the increase of the fire risk index calculated from air temperature, relative humidity, wind speed and drought factor; the probability of tree obstruction decreases with the increase of the clearance margin between the conductor and the vegetation corridor.

3. The main and distribution network coordinated scheduling method considering fault risk in mountainous areas according to claim 1, characterized in that, Step S3 includes: The relative deviation between the predicted failure frequency and the measured failure frequency of the failure probability model is calculated, and the Brier score between the predicted probability and the actual occurrence of the event is calculated. When the relative deviation and the Brier score do not exceed their respective set thresholds, the current model is adopted and marked as highly reliable; otherwise, the model parameters are updated recursively.

4. The main and distribution network coordinated scheduling method considering fault risk in mountainous areas according to claim 3, characterized in that, The model parameters include the tripping conversion coefficient of the lightning strike model, the regression coefficient of the wildfire model, and the vegetation growth rate and feature margin of the tree barrier model. The updated parameters are used to recalculate the verified fault probability.

5. The main and distribution network coordinated scheduling method considering fault risk in mountainous areas according to claim 1, characterized in that, Step S4 includes: Linearize the power flow equations at the current operating point to obtain the voltage-power sensitivity of each unit's injected power to the critical node voltage; The maximum injectable power of the unit under voltage constraints is obtained by dividing the difference between the node voltage limit and the current voltage of the critical node by the voltage-power sensitivity. The smaller value between the maximum injectable power and the rated adjustable capability of the unit is taken as the available adjustable capability of the unit; The difference is obtained by subtracting the sum of the available adjustment capabilities of each unit from the main network scheduling command. When the difference is greater than 0, the difference is the adjustment gap; when the difference is less than or equal to 0, the adjustment gap is 0.

6. The main and distribution network coordinated scheduling method considering fault risk in mountainous areas according to claim 1, characterized in that, Step S5 includes: Calculate the comprehensive risk score for each section; The feeder comprehensive risk index is calculated by summing the product of the comprehensive risk score of each section and the load power lost after the section is isolated by the sectionalizing switch on each section; and the feeder comprehensive risk index is divided into three levels: high, medium and low according to a set threshold.

7. A method for coordinated scheduling of main and distribution networks considering fault risks in mountainous areas, as described in claim 6, is characterized in that... The set threshold is determined based on the quantile of the historical power loss distribution of the feeder, or based on the power supply reliability target and the acceptable expected power loss level.

8. The main and distribution network coordinated scheduling method considering fault risk in mountainous areas according to claim 1, characterized in that, Step S6 includes: when the adjustment gap is equal to 0: The objective is to minimize the sum of the squares of the adjustment amounts of each adjustment unit after verification by the fault probability of its respective section, and the constraint is that the sum of the adjustment amounts of each unit is equal to the main grid dispatching instruction, and the adjustment amounts of each unit do not exceed the upper limit of the available adjustment capacity of the voltage constraint of that unit. When all constraints are not saturated, the solution is that each unit shares the adjustment demand proportionally according to the redistribution weight. The redistribution weight is the reciprocal of the sum of the fault probability of the section where the unit is located after verification and a regularization parameter. When the allocation of a certain unit reaches the upper limit of its voltage constraint available adjustment capability, the unit is fixed at the upper limit and deducted from the adjustment demand to be allocated. Then the remaining units are solved again, and the cycle is repeated until no unit exceeds the limit, and the corrected scheduling scheme is obtained.

9. A method for coordinated scheduling of main and distribution networks considering fault risks in mountainous areas, as described in claim 1, is characterized in that... Step S6 further includes: The feeder comprehensive risk index, the available regulation capacity of each regulation unit, and the regulation gap are reported to the main grid. Based on this, the main grid regenerates the next cycle dispatch instruction and issues it to the distribution network, forming an iterative collaboration between the main grid and the distribution network.

10. A main and distribution network coordinated dispatching system considering the fault risk in mountainous areas, characterized in that, The method applied to any one of claims 1 to 9 includes a data acquisition module, a fault probability modeling module, a meteorological verification and parameter update module, a tidal flow and regulation capacity assessment module, a comprehensive risk assessment module, and a deterministic scheduling correction and decision support module. The data acquisition module is used to collect meteorological data, lightning activity data, vegetation and line ledger data, and real-time operation data of the power distribution network. The failure probability modeling module is used to calculate the failure probability of each partition segment; The meteorological verification and parameter update module is used to calculate the relative deviation between the predicted fault frequency and the measured fault frequency or to calculate the Brier score between the predicted probability and the actual occurrence of the event, and compare it with the respective set thresholds to determine whether to perform an update verification. The power flow and regulation capacity assessment module is used to calculate the available regulation capacity of each regulation unit in the distribution network to obtain the regulation gap; The comprehensive risk assessment module is used to calculate the risk index and classification based on external environment data; The deterministic scheduling correction and decision support module is used to output the corrected scheduling scheme and risk warning.

Images